Drawing clustered graphs in three dimensions

Abstract

Clustered graph is a very useful model for drawing large and complex networks. This paper presents a new method for drawing clustered graphs in three dimensions. The method uses a divide and conquer approach. More specifically, it draws each cluster in a 2D plane to minimise occlusion and ease navigation. Then a 3D drawing of the whole graph is constructed by combining these 2D drawings. Our main contribution is to develop three linear time weighted tree drawing algorithms in three dimensions for clustered graph layout. Further, we have implemented a series of six different layouts for clustered graphs by combining three 3D tree layouts and two 2D graph layouts. The experimental results with metabolic pathways show that our method can produce a nice drawing of a clustered graph which clearly shows visual separation of the clusters, as well as highlighting the relationships between the clusters. Sample drawings are available from http://www.cs.usyd.edu.au/~visual/valacon/gallery/C3D/ © Springer-Verlag Berlin Heidelberg 2005.